Appendix A: Regular Expressions It’s All Greek to Me Regular Expressions • A pattern that matches a set of one or more strings • May be a simple string, or contain wildcard characters or modifiers • Used by programs such as vim, grep, awk, and sed • Not the same as shell expansion Components • Characters – Literals – Special Characters • Delimiters – Mark beginning end of regular expressions – Usually / – ’ (but not really) Simple Strings • Contain no special characters • Matches only the string • Ex: /foo/ matches: – foo – tomfoolery – bar.foo.com Special Characters • Can match multiple strings • Represent zero or more characters • Always match the longest possible string (we’ll see examples in a bit) Periods • Matches any single character • Ex: /.ing/ – I was talking – bling – he called ingred • Ex: /spar.ing/ – sparring – sparking Brackets • Define a character class • Match any one character in the class • If a carat (^) is first character in class, character class matches any character not in class • Other special characters in class lose meaning Brackets con’t • • • • Ex /[jJ]ustin/ matches justin and Justin Ex /[A-Za-z]/ matches any letter Ex /[0-9]/ matches any number Ex /[^a-z]/ matches anything but lowercase letters Asterisks • Zero or more occurrences of the previous character • So match any number of characters would be /.*/ • Ex /t.*ing/ – thing – this is really annoying Plus Signs and Question Marks • Very similar to asterisks, depend on previous • + matches one or more occurrences (not 0) • ? Matches zero or one occurrence (no more) • Ex /2+4?/ matches one or more 2’s followed by either zero or one – 22224, match – 4, 244 not • Part of the class of extended R.E Carets & Dollar Signs • If a regular expression starts with a ^, the string must be at the beginning of a line • If a regular expression ends with a $, the string must be at the end of a line • ^ and $ are referred to as anchors • Ex /^T.*T$/ matches any line that starts and ends with T Quoting Special Characters • If you want to use a special character literally, put a backslash in front of it • Ex /and\/or/ matches and/or • Ex /\\/ matches \ • Ex /\**/ matches any number of asterisks Longest Match • Regular expressions match the longest string possible in a line • Ex I (Justin) like coffee (lots) • /(.*)/ – Matches (Justin) like coffee (lots) • /([^)]*)/ – Matches (Justin) Boolean OR • You can pattern match for two distinct strings using OR (the pipe) • Ex /CAT|DOG/ – Matches exactly CAT and exactly DOG • Simplier expressions can be written just using a character class – I.E /a[bc]/ instead of /ab|ac/ • Also part of extended R.E Grouping • You can apply special characters to groups of characters in parenthesis • Also called bracketing • Matches same as unbracketed expression • But can use modifiers • Ex /\(duck\)*|\(goose\)/ Using with vim • Use regular expressions for searching and substituting • Searching: – /string or ?string • Substituting: – – – – :[g][address]s/string/replace[/g] g : global; substitute all lines string and replace can be R.E /g : global; replace all occurrences in the line Using with vim con’t • [address] – – – – – – n : line number n[+/-]x : line number plus x lines before or after n1,n2 : from line n1 to n2 : alias for current line $ : alias for last line in work buffer % : alias for entire work buffer vim examples • /^if( • /end\.$ • :%s/[Jj]ustin/Mr\ Awesome/g Using with vim con’t • Ampersand (&) – Alias for matched string when substituting – Ex: /[A-Z][0-9]/_&_/ • Quoted digit (\n) – Used with R.E with multiple quoted parts – Can be used to rearrange columns – Ex: /\([^,]*\), \(.*\)/\2 \1/ Using with grep • To take advantage of extended regular expressions, use egrep or grep -E instead • Use single quote as delimiter • Ex: – egrep ’^T.*T$’ myfile Lists all lines in myfile that begin & end with T ... Match any one character in the class • If a carat (^) is first character in class, character class matches any character not in class • Other special characters in class lose meaning Brackets... to as anchors • Ex /^T.*T$/ matches any line that starts and ends with T Quoting Special Characters • If you want to use a special character literally, put a backslash in front of it • Ex /and/or/... can be written just using a character class – I.E /a[ bc]/ instead of /ab|ac/ • Also part of extended R.E Grouping • You can apply special characters to groups of characters in parenthesis • Also